Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
tags:
|
6 |
+
- control-vector
|
7 |
+
- creative-writing
|
8 |
+
---
|
9 |
+
|
10 |
+
## gghfez/Sorcerer_creative-writing-control-vectors
|
11 |
+
|
12 |
+
Creative Writing control-vectors for [rAIfle/SorcererLM-8x22b-bf16](rAIfle/SorcererLM-8x22b-bf16)
|
13 |
+
|
14 |
+
This is an excellent model, but I found it somewhat diffcult to steer via prompting.
|
15 |
+
|
16 |
+
These control vector solved it for me.
|
17 |
+
|
18 |
+
## Usage example
|
19 |
+
|
20 |
+
```sh
|
21 |
+
llama-server --model SorcererLM-8x22b-iMat-IQ4_XS-00001-of-00002.gguf [other CLI arguments] \
|
22 |
+
--control-vector SorcererLM-8x22b-optimism_vs_nihilism__debias.gguf \
|
23 |
+
--control-vector <add as many as you want like this>.gguf
|
24 |
+
```
|
25 |
+
|
26 |
+
Train on 1xH100NVL using jukofyork's [control-vectors](https://github.com/jukofyork/control-vectors)
|
27 |
+
|
28 |
+
Based off Version v3.0 of [creative-writing-control-vectors-v3.0](jukofyork/creative-writing-control-vectors-v3.0]).
|
29 |
+
|
30 |
+
See here for more info about control-vectors [creative-writing-control-vectors-v3.0](jukofyork/creative-writing-control-vectors-v3.0])
|